Unlock instant, AI-driven research and patent intelligence for your innovation.

Dataflow component scheduling using reader/writer semantics

a dataflow and component scheduling technology, applied in multi-programming arrangements, program control, instruments, etc., can solve problems such as data race, unrestricted access to shared mutable states,

Active Publication Date: 2011-06-30
MICROSOFT TECH LICENSING LLC
View PDF5 Cites 16 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, unrestricted access to shared mutable state is prone to a data race, a situation where a node can observe an invalid state of an object in memory, while another node is at the same time mutating the state of that object.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Dataflow component scheduling using reader/writer semantics
  • Dataflow component scheduling using reader/writer semantics
  • Dataflow component scheduling using reader/writer semantics

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0014]In accordance with embodiments described herein, the scheduling of dataflow components in a dataflow network is described. A number, if not all, of the dataflow components are created using a domain / agent model. A scheduler identifies, for a number of the components, a creation source for the given component. The schedule also identifies an appropriate domain-level access permission (and potentially also an appropriate agent-level access permission) for the given component based on the creation source of the given component. Tokens may be used at the domain or agent level to control access. First, some introductory discussion regarding computing systems will be described with respect to FIG. 1. Then, various embodiments of the scheduling of dataflow components will be described with reference to FIGS. 2 through 7.

[0015]First, introductory discussion regarding computing systems is described with respect to FIG. 1. Computing systems are now increasingly taking a wide variety of ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The scheduling of dataflow components in a dataflow network. A number, if not all, of the dataflow components are created using a domain / agent model. A scheduler identifies, for a number of the components, a creation source for the given component. The scheduler also identifies an appropriate domain-level access permission (and potentially also an appropriate agent-level access permission) for the given component based on the creation source of the given component. Tokens may be used at the domain or agent level to control access.

Description

BACKGROUND[0001]Computing systems obtain a high degree of functionality by executing software programs. A dataflow network is a directed graph where nodes represent computations and the edges represent data flowing between the nodes. Dataflow networks are ubiquitous in software. They are relatively easy to build, reason about, and provide a degree of parallelism by allowing multiple nodes to execute in parallel.[0002]In typical dataflow systems, execution of the nodes of the dataflow network is driven solely by the availability of the input expected by the nodes. Isolated and side-effect free dataflow networks (such as Kahn process networks) do not read or alter the external state. Conversely, non-isolated dataflow networks (such as Unix pipes) can read and / or alter the external state. Non-isolated networks have a potential of better performance due to their ability to perform in-place updates and avoid data copying. However, unrestricted access to shared mutable state is prone to a...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(United States)
IPC IPC(8): G06F9/46
CPCG06F9/52
Inventor LAKSBERG, ARTURPHILLIPS, JOSHUA D.GUSTAFSSON, NIKLAS
Owner MICROSOFT TECH LICENSING LLC